No description available.

This thesis proposes a novel theory to account for the structural Case facts in Standard Arabic (SA). It argues that structural Nom and Acc Cases are licensed by Verbal Case (VC). Thus it argues against the proposal that structural Case in SA is licensed as a reflex of phi-agreement (Schütze 1997 and Chomsky 2001 crosslinguistically, and Soltan 2007 for SA), and also against the view that structural Case is a [uT] feature on the DP (Pesetsky & Torrego 2001, 2004). After arguing against these two approaches, it is shown that verbless sentences, where the verb is not licensed (by VC), do not witness the licensing of structural Case. Thus verbless sentences provide a context where verbs are not licensed, similar to the embedded subject position of control verbs like ‘try’ (where lexical DPs are not licensed). Investigation of the SA verbal system reveals that SA verbs are licensed through Case checking/assignment by verbal particles. Thus, like DPs, verbs receive a form of Case, which I call VC, represented as unvalued [VC] features on I and v*. Since the VC-assigning particles are Comp elements, I propose that [VC] is valued on I and v* by a valued [VC] feature on Fin (via Agree), which enables I and v* to value the [Case] features on the subject and object as Nom and Acc, respectively. Thus the DP is licensed by the same feature that licenses the verb, which is VC. Given the observation that [T], [phi], and [Mood] do not license Case in SA, I argue for two types of finiteness, Infl-finiteness, related to [T], [Mood], and [phi], and Comp-finiteness, related to [VC]. To account for the Case facts in various SA sentence types, I propose that Fin0 has a [VC] feature iff it selects an XP that has (at least) one I-finiteness feature ([T], [Mood], [phi]) and a categorial [V] feature.

This thesis proposes a novel theory to account for the structural Case facts in Standard Arabic (SA). It argues that structural Nom and Acc Cases are licensed by Verbal Case (VC). Thus it argues against the proposal that structural Case in SA is licensed as a reflex of phi-agreement (Schütze 1997 and Chomsky 2001 crosslinguistically, and Soltan 2007 for SA), and also against the view that structural Case is a [uT] feature on the DP (Pesetsky & Torrego 2001, 2004). After arguing against these two approaches, it is shown that verbless sentences, where the verb is not licensed (by VC), do not witness the licensing of structural Case. Thus verbless sentences provide a context where verbs are not licensed, similar to the embedded subject position of control verbs like ‘try’ (where lexical DPs are not licensed). Investigation of the SA verbal system reveals that SA verbs are licensed through Case checking/assignment by verbal particles. Thus, like DPs, verbs receive a form of Case, which I call VC, represented as unvalued [VC] features on I and v*. Since the VC-assigning particles are Comp elements, I propose that [VC] is valued on I and v* by a valued [VC] feature on Fin (via Agree), which enables I and v* to value the [Case] features on the subject and object as Nom and Acc, respectively. Thus the DP is licensed by the same feature that licenses the verb, which is VC. Given the observation that [T], [phi], and [Mood] do not license Case in SA, I argue for two types of finiteness, Infl-finiteness, related to [T], [Mood], and [phi], and Comp-finiteness, related to [VC]. To account for the Case facts in various SA sentence types, I propose that Fin0 has a [VC] feature iff it selects an XP that has (at least) one I-finiteness feature ([T], [Mood], [phi]) and a categorial [V] feature.

abstract: This dissertation provides an account of evidentiality of a number of selected verbs of perception in Modern Standard Arabic (MSA). The verbs are divided into three categories: activity, experiential, and source-based, following Viberg (1983). The data shows that the activity P.Vs in MSA are rarely used evidentially whereas the experiential and the source-based ones are commonly used to indicate evidential meaning. It also shows that while the source-based verb is mostly used with an inferred evidential meaning, the evidentiality encoded by the experiential perception verbs is determined by the complementation pattern and the person of the subject (first or third person subject). With the non-finite complement, these verbs indicate a direct evidentiality when having a first person subject, and a reported evidentiality when having a third person subject. With the finite CP complement, they indicate an indirect evidentiality. This corpus-based study also examines the grammaticalization of these verbs when used evidentially. I argue that only the verb ra’aa of the selected experiential verbs is fully grammaticalized, but only when it is in the past tense and followed by a verbal non-finite complement. In this usage, it becomes a light verb. The source-based verb badaa/yabduu when indicating an evidentiality, it is grammaticalized into copulative verb when followed by an adjectival predicate, and modal verb when followed by a finite complement. / Dissertation/Thesis / Doctoral Dissertation English 2019

Research for automatic readability prediction of text has increased in the last decade and has shown that various machine learning methods can effectively address this problem. Many researchers have applied machine learning to readability prediction for English, while Modern Standard Arabic (MSA) has received little attention. Here I describe a system which leverages machine learning to automatically predict the readability of MSA. I gathered a corpus comprising 179 documents that were annotated with the Interagency Language Roundtable (ILR) levels. Then, I extracted lexical and discourse features from each document. Finally, I applied the Tilburg Memory-Based Learning (TiMBL) machine learning system to read these features and predict the ILR level of each document using 10-fold cross validation for both 3-level and 5-level classification tasks and an 80/20 division for a 5-level classification task. I measured performance using the F-score. For 3-level and 5-level classifications my system achieved F-scores of 0.719 and 0.519 respectively. I discuss the implication of these results and the possibility of future development.

abstract: This thesis investigates the pronominal system in Standard Arabic. It seeks to unravel the correlation between independent and dependent personal pronouns. Although both pronoun groups are treated as distinct parts of the lexicon, I argue that dependent pronouns are reduced forms derived from the strong counterparts. This study examines how these forms (reduced and non-reduced) relate to one another phonologically and syntactically. Various analytical tools are utilized including vowel harmony, syllable structure as well as some principles of Distributed Morphology and Chomsky's 1995 Minimalist Program. With regard to the phonological relations, I argue that dependent subject pronouns are generated from their parallel strong forms by omitting the initial syllable. Dependent object pronouns are formed by omitting the first two syllables. The first person singular and third person plural masculine subject pronouns are suppletive forms completing the paradigm. They are not derived by reduction from their full counterparts. After investigating the distributional properties of both sets of pronouns, I propose a bipartite subcategorization of reduced pronominals into two subclasses: clitics and affixes. Clitics surface in positions in which strong pronouns cannot occur. As for affixes, they are used to mark verb-argument agreement. In light of these positions, I argue that dependent subject pronouns are always affixes while dependent object pronouns are always clitics. Clitics function as syntactically independent units which combine with hosts at the phonological phase as a result of their prosodic deficiency while affixes associate with hosts when features are valued during a sentence derivation. / Dissertation/Thesis / M.A. English 2013

This thesis examines the difference between conditional phrases in Egyptian Colloquial (EC) and Modern Standard Arabic (MSA). It focuses on two different conditional particles 'iḏa and law. Verb tenses featured after the conditional particle determine the difference between EC and MSA usage. Grammars for EC and MSA provide a prescriptive approach for a comparison with empirical data from Arabic corpora. The study uses data from the ArabiCorpus along with a corpus of Egyptian Colloquial that were compiled specifically for this study. The results of this study demonstrate that each particle (‘iḏa and law) and register (EC and MSA) favors a certain tense. Also, the data contrast with rules prescribed by grammars for MSA. Present tense verbs appear in the proposed condition for particle law a total of 22 out of 400 tokens (5.5%). Verb tense also plays an important role in determining the connecting particle for MSA sentences. The results demonstrate that the selection of connecting particles for law does not occur by chance but is instead systematic in nature. An apodosis containing a past tense verb strongly favors the connecter la, while one with a non-past tense verb strongly favors the connector fa.

abstract: ABSTRACT This dissertation investigates the copular/locative and existential predications in Arabic. The main focus is on the typology and syntax of the existential predications. The negation of such predications reveals interesting results. The Negative Existential Cycle (Croft, 1991) is a model that describes the process by which verbal negators arise from existential negators. I discuss data of existentials and negative existentials from Standard Arabic, Saudi Arabic dialect, and Gulf Pidgin Arabic. I argue for canonical vs. non-canonical word orders in copular/locative and existential sentences, respectively. I examine the grammaticalization path of the existentials from their locative content in each language form. Then, I investigate the syntactic word order of the copular/locative and existential constructions in each variety. I investigate the negation of the existential construction in each variety. First, Standard Arabic is shown to be at stage A in the Negative Existential Cycle. The Hijazi and Najdi Arabic spoken by elders show further developments. Hijazi Arabic appears to be at stage B, while Najdi Arabic appears to be at stage B and an intermediate stage B ~ C. Second, I show that in Saudi Arabic the negative existential has been extended to the verbal domain. Saudi Arabic is at stages A, B, and B ~ C, while Qassimi Arabic is at stages A and B. Third, I show that the existential construction in Gulf Pidgin Arabic is only negated by the negative existential predicate, while the verbal sentences are negated by the negative existential and the verbal negator. Therefore, Gulf Pidgin Arabic is at stages B and C in the Negative Existential Cycle. Finally, I discuss the syntax of copular/locative and existential predications in each variety. I propose a unified syntactic structure. Existential and possessive predications are analyzed as inverse copular sentences (Moro, 1997) as opposed to the canonical copular/locative sentences. The unified structure accounts for the agreement facts, such as partial vs. full agreement in existential and copular/locative predications, respectively. The data investigated here will contribute to Arabic comparative and historical linguistics. More Arabic dialects’ data is needed to determine their stages in the Negative Existential Cycle. / Dissertation/Thesis / Doctoral Dissertation English 2019

Die vorliegende Arbeit soll beispielhaft die Verwendung von dialektalen und hochsprachlichen Elementen im gesprochenen Arabisch untersuchen. Dabei erhebt sie jedoch keinen dialektologischen Anspruch, sondern das Hauptaugenmerk liegt vielmehr auf der interdialektalen Kommunikation gebildeter Sprecher. Im weitesten Sinne soll diese Arbeit einen Beitrag dazu zu leisten, die Frage nach einer dialektübergreifenden, gehobenen arabischen Umgangssprache zu beantworten. Bei der Wahl des Themas war die Überlegung ausschlaggebend, dass Studenten der arabischen Sprache und insbesondere des Faches Dolmetschen in ihrem Studium ausschließlich die moderne arabische Schriftsprache vermittelt wird. Außerhalb der Lehrveranstaltungen werden sie jedoch bald mit den arabischen Dialekten konfrontiert, die die eigentliche Muttersprache arabischer Sprecher darstellen. Aus diesem Grund kommen sie nicht umhin, sich zumindest passive Sprachkenntnisse in einigen dieser Dialekte anzueignen, um nicht nur Nachrichtensendungen auf Hocharabisch folgen zu können, sondern auch mit Muttersprachlern aus verschiedenen Regionen der arabischen Welt kommunizieren zu können. Besonders für Studierende des Faches Dolmetschen ist es im Hinblick auf ihre spätere Tätigkeit nicht ratsam, sich auf eine bestimmte Region festzulegen. Zudem sind für Dolmetscher in erster Linie fachliche Themen, etwa aus den Bereichen Politik und Wirtschaft, und vergleichsweise formelle Anlässe von Interesse. Für die Bearbeitung des Themas wurde daher eine Fernsehsendung ausgewählt, die nicht nur die Möglichkeit bietet, die Kommunikation zwischen Sprechern mit unterschiedlichem dialektalem Hintergrund zu untersuchen, sondern auch Themengebiete behandelt, die für die spätere Dolmetschertätigkeit der Studierenden relevant sind. Die Wahl der politischen Talkshow Ḥiwār al-ᶜArab, in der Themen von öffentlichem Interesse von Fachleuten aus verschiedenen arabischen Ländern diskutiert werden, bot sich daher an. Um den Rahmen einer Studienabschlussarbeit nicht zu sprengen, wurde nur eine Ausgabe der genannten Sendung für die sprachwissenschaftliche Analyse herangezogen. Die Redebeiträge der verschiedenen Sprecher, die in der untersuchten Ausgabe auftreten, sollen auf dialektale und hochsprachliche Elemente untersucht werden, um so Arabischlernenden einen Einblick zu verschaffen, welche Arten von Dialektinterferenzen in realen Kommunikationssituationen auftreten können. Nach Möglichkeit soll zudem bereits eine erste grobe Einteilung verschiedener Stufen von Dialektinterferenzen vorgenommen werden. Dabei konnten jedoch nicht alle auftretenden sprachlichen Merkmale berücksichtigt werden. Aus diesem Grund wurden für jeden Sprecher nur die jeweils wichtigsten phonetischen, lexikalischen und grammatischen Eigenheiten beschrieben. Auch wurden die Studenten, die sich nur sehr kurz in der Sendung äußern, bei der Analyse außer Acht gelassen. Da für die Talkshow Ḥiwār al-ᶜArab keine schriftliche Fassung vorliegt, wurde die untersuchte Ausgabe der Sendung nach Gehör transkribiert. Dabei wurde versucht, der tatsächlichen Aussprache so weit wie möglich gerecht zu werden. Um die Transkription auch ohne Kenntnis der sprachlichen Analyse so verständlich wie möglich zu halten, wurden jedoch nicht alle beobachteten phonetischen Merkmale in der Umschrift wiedergegeben. Gegebenenfalls wurden bestimmte Laute in der Analyse näher beschrieben. Die Wiedergabe in IPA-Transkription steht dabei in eckigen Klammern. Inhaltlich wurden lediglich sehr schlecht hörbare Satzteile und einzelne vom Moderator eingeworfene Wörter ohne besondere Bedeutung („Ṭayyib, ṭayyib…“) ausgelassen, die nicht zum Ziel haben, den Redenden zu unterbrechen. Satzzeichen wurden nach eigenem Ermessen und Intonation des Sprechers gesetzt und dienen lediglich der besseren Lesbarkeit.

The aim of this thesis is to investigate a number of factors that could affect the performance of an Arabic automatic speech understanding (ASU) system. The work described in this thesis belongs to the speech recognition (ASR) phase, but the fact that it is part of an ASU project rather than a stand-alone piece of work on ASR influences the way in which it will be carried out. Our main concern in this work is to determine the best way to exploit the phonological properties of the Arabic language in order to improve the performance of the speech recogniser. One of the main challenges facing the processing of Arabic is the effect of the local context, which induces changes in the phonetic representation of a given text, thereby causing the recognition engine to misclassifiy it. The proposed solution is to develop a set of language-dependent grapheme-to-allophone rules that can predict such allophonic variations and eventually provide a phonetic transcription that is sensitive to the local context for the ASR system. The novel aspect of this method is that the pronunciation of each word is extracted directly from a context-sensitive phonetic transcription rather than a predened dictionary that typically does not reect the actual pronunciation of the word. Besides investigating the boundary effect on pronunciation, the research also seeks to address the problem of Arabic's complex morphology. Two solutions are proposed to tackle this problem, namely, using underspecified phonetic transcription to build the system, and using phonemes instead of words to build the hidden markov models (HMMS). The research also seeks to investigate several technical settings that might have an effect on the system's performance. These include training on the sub-population to minimise the variation caused by training on the main undifferentiated population, as well as investigating the correlation between training size and performance of the ASR system.